Archives of Sexual Behavior https://doi.org/10.1007/s10508-018-1194-7

ORIGINAL PAPER

Heterosexual College Students Who Hookup with Same‑Sex Partners

Arielle Kuperberg1 · Alicia M. Walker2

Received: 9 December 2016 / Revised: 13 March 2018 / Accepted: 14 March 2018 © Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract Individuals who identify as heterosexual but engage in same-sex sexual behavior fascinate both researchers and the media. We analyzed the Online College Social Life Survey dataset of over 24,000 undergraduate students to examine students whose last hookup was with a same-sex partner (N = 383 men and 312 women). The characteristics of a signifcant minority of these students (12% of men and 25% of women) who labelled their sexual orientation “heterosexual” difered from those who self- identifed as “homosexual,” “bisexual,” or “uncertain.” Diferences among those who identifed as heterosexual included more conservative attitudes, less prior homosexual and more prior heterosexual sexual experience, features of the hookups, and sentiments about the encounter after the fact. Latent class analysis revealed six distinctive “types” of heterosexually identifed students whose last hookup was with a same-sex partner. Three types, comprising 60% of students, could be clas- sifed as mostly private sexual experimentation among those with little prior same-sex experience, including some who did not enjoy the encounter; the other two types in this group enjoyed the encounter, but difered on drunkenness and desire for a future relationship with their partner. Roughly, 12% could be classifed as conforming to a “performative bisexuality” script of women publicly engaging in same-sex hookups at college parties, and the remaining 28% had strong religious practices and/or beliefs that may preclude a non-heterosexual identity, including 7% who exhibited “internalized heterosexism.” Results indicate several distinctive motivations for a heterosexual identity among those who hooked up with same-sex partners; previ- ous research focusing on selective “types” excludes many exhibiting this discordance.

Keywords Hookups · Same-sex sexual behavior · Sexual identity · Internalized heterosexism

Introduction when given that option, and same-sex attraction was reported by 5% of men and 13% of women, but only 2% of men and Many people engage in same-sex sexual encounters or desire 4% of women identifed as LGB (Savin-Williams & Ream, them but maintain a heterosexual identity; others who adopt a 2007). Another study representative of U.S. 15–44 year olds lesbian, gay, bisexual (LGB) or other sexual minority identity found 9% of women and 3% of men who identifed as het- later relinquish it for a heterosexual one (Diamond, 2003; erosexual had same-sex sexual experience (Chandra, Copen, Hamilton, 2007; Walker, 2014a, 2014b; Ward, 2015). One & Mosher, 2013). study of college students found 30% of women and 19% of Several theoretical frameworks have been proposed to men who identifed as heterosexual reported same-sex attrac- explain this discordance. Some studies with limited samples tion (Hoburg, Konik, Williams, & Crawford, 2004). A study examined “the closet” or “the down low” (Boykin, 2005; Ford, representative of 18–26 year olds in the U.S. found 3% of Whetten, Hall, Kaufman, & Thrasher, 2007; King, 2004; Phil- men and 11% of women identifed as “mostly heterosexual” lips, 2005). Others focused on college “hookup culture” and expectations of sexual experimentation, including young women hooking up with other women at parties, ostensibly to * Arielle Kuperberg attract men (Diamond, 2005; Kimmel & Plante, 2002; Kuper- [email protected] berg & Padgett, 2015; Wade, 2017; Ward, 2015). A third line 1 Sociology Department, The University of North Carolina of research examined LGB identity acquisition (Cass, 1979, at Greensboro, PO Box 26170, Greensboro, NC 27402‑6170, 1996; Horowitz & Newcomb, 2002; Kaufman & Johnson, USA 2004). Negative feelings about homosexuality among those 2 Department of Sociology and Anthropology, Missouri State with same-sex attractions or “internalized heterosexism” may University, Springfeld, MO, USA

Vol.:(0123456789)1 3 Archives of Sexual Behavior also be a factor (Kaufman & Johnson, 2004; Taylor, 1999). We sexual encounters, they rejected a bisexual identity, instead examined whether a heterosexual identity among these students explaining these acts as due to their “freakiness,” which they correlated with characteristics that would be predicted by these felt was the accurate term for their sexual orientation (Walker, distinct and sometimes-competing frameworks, and the extent 2014b). Budnick (2016) found women with the least edu- to which students comprised distinguishable groups described cation reported the most lifetime same-sex sexual events, by these theories. but for some early entry into motherhood closed-of sexual exploration and possible development of a LGBQ identity. Same‑Sex Hookups Among Self‑Identifed Heterosexuals College Hookup Scripts and Same‑Sex Hookups Among Heterosexuals Public fascination with self-identifed heterosexuals hooking up with same-sex partners arose in the 1990s, along with phrases Recent literature examined college hookups, casual sexual “on the down low” and “on the D.L.” Originating among Afri- encounters which most college students participate in, that can-Americans, these idioms originally referred to any act done can range from an intense “make-out” session to intercourse secretly, but became associated with men who “publicly pre- (Bogle, 2008; England, Shafer, & Fogarty, 2008; Kuperberg sent as heterosexual while secretly having sex with other men” & Padgett, 2015, 2016; Reiber & Garcia, 2010). Hookup (Boykin, 2005; Ford et al., 2007; Phillips, 2005). Others used rates and risk-taking during hookups have been found to dif- “in the closet” to describe LGB individuals hiding their sexual fer by gender, GPA, race, religiosity, mother’s education, and identity in public, or even to themselves (Seidman, Meeks, & age (Kuperberg & Padgett, 2015, 2016, 2017). The college Traschen, 1999). Academic literature opted for “men who have hookup scene is an opportunity for students to experiment sex with men” or “MSM,” a term potentially obscuring the with and afrm non-heterosexual sexual identities or to con- meaning-making of sexuality (Young & Meyer, 2005). frm a heterosexual one (Rupp, Taylor, Regev-Messalem, Research on same-sex sexual encounters is generally lim- Fogarty, & England 2013). Social “scripts,” or expectations ited and subject to sample bias. Some examined risk-taking of behavior, position college as a “time to experiment” sexu- and sexual activity in same-sex hookups from a larger sample, ally (Kuperberg & Padgett, 2015; Simon & Gagnon, 2003) but most research focused on select groups, such as couples in and may encourage same-sex hookups even among those long-term relationships or those identifying as LGB (Eisenberg, without same-sex attractions. Many dismiss these hookups 2001; Kuperberg & Padgett, 2015, 2017; Peplau & Fingerhut, as “experimentation” or “accidental” (Ward, 2015). 2007; Rust, 1992). Past research also mostly focused on men Some women engage in same-sex hookups to attract men’s and recruited subjects from biased sources, including web- attention as an established part of the college hookup “sexual sites, bars, or parks known to be frequented by MSM or LGBT script” (Diamond, 2005; Wade, 2017). Women often conduct organizations and magazines (Brady & Busse, 1994; CDC, this “performative bisexuality” in public spaces for the bene- 2010; Cole, Kemeny, Taylor, & Visscher, 1996; Hightow et al., ft of a male audience, as it is a commonly reported “turn-on” 2006; Kaufman & Johnson, 2004; Koblin et al., 2000; Lindley, for heterosexual men (Kimmel & Plante, 2002). Hamilton Nicholson, Kerby, & Lu, 2003; Rhodes, DiClemente, Cecil, (2007) described a campus culture where heterosexually- Hergenrather, & Yee, 2002; Rosario, Schrimshaw, Hunter, & identifed women engaged in same-sex erotic behavior during Braun, 2006; Rowen & Malcolm, 2002). Much research on parties, later posting the pictorial evidence to social media. self-identifed heterosexuals hooking up with same-sex partners Often attributed to alcohol consumption, these encounters focused on African-American men, despite White men more included kissing and fondling breasts or buttocks, but no commonly exhibiting this disparity (Bleich & Taylor-Clark, genital contact, and allowed young women to mark them- 2005; Ford et al., 2007; King, 2004; Ross, Essien, Wiliams, & selves as “edgy” (Hamilton, 2007). Fernandez-Esquer, 2003; Ward, 2015). This focus may stem However, suggesting that women engage in this behavior only from perceptions that “the down low” is limited to African- to attract men may obscure some functions of these encoun- American men due to the origins of the term or heightened ters. In one study, two female roommates reported dancing surveillance of the sexuality of men of color (Ward, 2015). naked together when they were alone in their room, but jok- Women also seek out same-sex sexual partners while iden- ingly dismissed it as an activity for “when they were bored,” tifying as heterosexual (Walker, 2014a, b), although research while another study found women dismissing private same-sex on students included only small samples of 80 or fewer (Dia- hookups as a result of inebriation (Hamilton, 2007; Wade, 2017). mond, 2003; Hamilton, 2007; Peterson & Gerrity, 2006), and Hooking up with other women at alcohol-fueled public parties research outside of college contexts is even more limited. allowed women, particularly those who had negative views of Walker (2014a) found women married to men who had afairs lesbianism, to experiment with socially acceptable same-sex with women felt these encounters “didn’t count” in terms of behavior assumed to be intended for male pleasure (Hamilton, monogamy. While reporting lifelong same-sex attraction and 2007; Wade, 2017; Ward, 2015).

1 3 Archives of Sexual Behavior

Other research focused on college men having sexual con- with other LGB individuals; (5) developing pride for identity, tact with men in “heterosexual contexts,” per established social perhaps anger toward society and heterosexuals; and (6) syn- scripts, including fraternity hazing rituals culturally defned as thesizing an LGB identity with other aspects of identity (Cass, heterosexual bonding activities rather than homosexual acts 1996). During early stages, individuals may attempt to recon- (Sanday, 2007; Silva, 2017; Simon & Gagnon, 1986, 2003; cile a heterosexual identity with same-sex sexual behavior and Ward, 2015). Members of organizations present these as ritu- attractions by interpreting them as temporary or a “special case” als of domination and humiliation aimed at increasing male (“If not for this special person whom I love, I would be hetero- bonding, and frame participants as having “no choice but to sexual”) (Cass, 1996). They may later adopt a LGB identity or comply,” creating a context for homosexual encounters contex- may cease that behavior and never adopt an LGB identity (Brady tualized as “obviously not gay,” while permitting sexual fex- & Busse, 1994; Cass, 1996; Horowitz & Newcomb, 2002; Kauf- ibility and experimentation among men (Ward, 2015). These man & Johnson, 2004). They may be in “transition,” where they hazing encounters and other encounters described as “situ- have begun to “recognize that they are not heterosexual, yet ational” homosexuality among men in specifc contexts (such have not adopted a homosexual identity” (Taylor, 1999). One as in prisons) are not presumed to be signifcant to underlying study of college women found of those identifed as being in one sexuality (Kimmel, 2008; Ward, 2015). of the stages of homosexual identity development, those with “heterosexual” identities were all in Stages 1–3 of the updated Sexual Identity Theory model (Peterson & Gerrity, 2006). Researchers critiqued the developmental stages model as While questions of sexual orientation aim to categorize sexual stemming from an essentialist perspective, with sexual identity attractions, desires, and behaviors, disparities between reported conceptualized as unchanging (Horowitz & Newcomb, 2002; orientations and behavior suggest these questions instead meas- Kaufman & Johnson, 2004; Peterson & Gerrity, 2006). Social ure sexual identity. Identity theory understands identities as a constructionists instead conceptualized identity as fuid over set of meanings individuals use to self-defne what it means to time and social context, infuenced by interactions that socially be in a particular role or situation, but which are malleable and create and reinforce that identity, and personal and social sig- can change over an individual’s lifetime (Rupp et al., 2013). nifcance at a specifc time and place (Horowitz & Newcomb, Commitment to, acceptance, and integration of an LGB identity 2002). Indeed, the idea that same-sex sexual behaviors con- is an ongoing process often lasting through adolescence and stitute an “identity” only fully emerged in the mid-twentieth beyond, with many frst adopting a bisexual identity before later century (Ward, 2015). Research found that the gender of indi- adopting a gay or lesbian identity (Rosario et al., 2006). Indi- vidual’s sexual interest can shift over their lifespan or in cer- viduals sometimes adopt identities to represent current sexual tain contexts (Baumeister, 2000; Blumstein & Schwartz, 1977; partnerships or choices, rather than to embody their overall feel- Diamond, 2003, 2008; Goode & Haber, 1997; Seidman et al., ings of attraction to members of either gender over their lifetime 1999; Sophie, 1986). LGB identities can develop in response (Seidman et al., 1999). Some lesbian and bisexual women later to positive self-appraisals, appraisals from others, or in the con- relinquish their sexual minority identity for a heterosexual one text of same-sex romantic relationships (Kaufman & Johnson, after forming relationships with men, to reconcile their identity 2004). Some LGB individuals who self-label receive negative and behavior (Diamond, 2003). Others describe their identity as appraisals from others and may deemphasize that identity to “heterofexible,” meaning they are mostly attracted to men, but avoid stigma, especially if they internalize those views (Kauf- occasionally participate in same-sex sexual behaviors they may man & Johnson, 2004; Taylor, 1999). describe as random, accidental, or meaningless (Ward, 2015). Along with biphobia, this “internalized heterosexism” may The “developmental stages model” theory of sexual identity prevent some from taking on a LGB identity (Dworkin, 2001; positioned taking on a homosexual identity and integrating it Hutchins & Kaahumanu, 1991; Ochs & Deihl, 1992; Peterson into your broader personal identity as the fnal stage in becom- & Gerrity, 2006; Rowen & Malcolm, 2002). Also called inter- ing aware of one’s underlying or “real” sexual orientation (Cass, nalized homophobia, the internalization of negative societal 1979, 1996; Horowitz & Newcomb, 2002; Kaufman & Johnson, attitudes toward homosexuality, or “heterosexism,” by those 2004). Researchers initially described the stages as (1) feelings with same-sex attractions is correlated with higher religiosity, of homosexual attraction and identity confusion; (2) homo- substance use, sexual risk-taking, and poor mental and physi- sexual experiences; (3) disclosing identities to some; and (4) cal health (Amadio, 2006; Kashubeck-West & Szymanski, sexual identity fully integrated into broader identity (Kaufman 2008; Rowen & Malcolm, 2002; Szymanski, Kashubeck- & Johnson, 2004). An updated model later took the focus away West, & Meyer, 2008). Most prominent religious groups in from homosexual experiences, with stages including (1) identity the USA oppose same-sex relationships, and internalization confusion; (2) evaluating familial and social consequences of of these views may explain why religious individuals who an LGB identity; (3) beginning to tolerate an LGB identity; (4) engage in same-sex behavior are more likely to identify as acceptance of identity, identifying to others, increasing contact heterosexual (Szymanski et al., 2008). Additionally, the term

1 3 Archives of Sexual Behavior

“bisexual” often meets resistance and may be avoided due Theory related to internalized heterosexism suggests hetero- to disbelief of bisexuality as a legitimate sexual orientation sexually identifed students may be more religious and socially (Israel & Mohr, 2004; Rupp et al., 2013; Yost & Thomas, conservative, have more negative sentiments about homosexu- 2012; Zivony & Lobel, 2014). Many bisexual individuals ality, and take more risks, such as unprotected sex and binge wonder if they are “bisexual enough” to warrant the identity drinking. We next ask: (Bower, Gurevich, & Mathieson, 2002; Ochs, 2007), while some privately consider themselves bisexual, but avoid social R3 Do these characteristics students support an internalized confict and rejection by allowing others to assume they are heterosexism model? heterosexual (Ochs & Deihl, 1992). We also explored racial patterns and how well students knew Present Study Objectives their partner to examine the degree to which Black men who have sex with anonymous male partners—the subject of much We analyzed the Online College Social Life Survey (OCSLS), prior study—are prevalent in these data. We also explored a dataset of over 24,000 students at 22 colleges and universi- whether some of these encounters were due to sexual assault. ties, to examine those who identifed as heterosexual, but These research questions included: whose last hookup partner was same sex. Using only data on students whose last hookup partner was same sex, we R4 Are Black students and those who do not know their analyzed whether a “heterosexual” identity correlated with sexual partners well more prevalent among men who identify demographics, attitudes, past sexual and relationship expe- as heterosexual, but hookup with same-sex partners? and rience, and hookup encounter characteristics, and whether students comprised distinct groups. College experimenta- R5 Were same-sex hookups among heterosexual students tion scripts and theories related to performative bisexuality the result of sexual assault? and fraternity hazing rituals suggest that same-sex encoun- ters among students who identify as heterosexual may be Prior theories may describe distinct groups that together more likely to take place among students with little prior comprise the students who identify as heterosexual but hookup same-sex experience, among fraternity and sorority mem- with same-sex partners. Prior research tended to examine one bers, those who are more accepting of or interested in sexual group at a time, such as women who engage in public hookups experimentation, involve only “low-level” sexual behavior, with women, or those who exhibit internalized heterosexism, take place in public social settings, and involve intoxication. but has not addressed the prevalence of various groups or the These students may be less likely than those with other iden- extent to which the wider group of college students who hookup tities to enjoy the encounter or want additional hookups or with same-sex partners can be described by these various and a relationship with their same-sex hookup partner. Our frst sometimes-competing theoretical frameworks. We draw upon research question asks: latent class analysis methods to examine a fnal central research question: Research Question(R)1 Do characteristics of same-sex hookups among heterosexual college students align with R6 Are college students who hookup with same-sex partners, college sexual scripting theory related to sexual experimen- but identify as heterosexual comprised of distinct “types” that tation, performative bisexuality, and/or fraternity hazing conform to various prevailing theories described in R1–R3, rituals? and what is the prevalence of each “type?”

The development stages model and the related social con- structionist model predict that students with a heterosexual Method identity would have fewer same-sex sexual experiences and more other-sex sexual experiences shaping their identity. The Subjects development stages model also suggests those with a hetero- sexual identity may be younger, and especially attracted to and We analyzed the OCSLS, a survey of romantic and sexual want a relationship with their partner (the “special case”). We partnering behavior collected between 2005 and 2011 from next ask: 24,131 students attending 22 colleges and universities. Ques- tions asked about students’ most recent dates and hookups, life- R2 Do these characteristics align with the developmental time sexual behavior, and a variety of demographic and attitude stages model and/or social construction model of identity questions. Professors of large introductory level courses and theory? courses addressing sociology, family, sexuality, gender, and

1 3 Archives of Sexual Behavior public health at these universities distributed surveys to students as “heterosexually-identifed” and the former as “non-hetero- as a course assignment, ofering an alternative assignment for sexually-identifed” for brevity; however, all such references students who did not participate. This sampling method resulted refer only to students whose most recent hookup was with a in a non-representative sample; elite research-oriented univer- same-sex partner. sities, underclasssmen, and women were overrepresented, and While the overall measure capturing same-sex hookups although almost 90% of participants were not sociology majors, did not include transgender partners, in terms of past experi- around 80% of courses in which data were collected were soci- ences with partners we included some transgender partners; ology courses (Kuperberg & Padgett, 2015). The response rate among non-heterosexually identifed students, two women’s was over 99% (Armstrong, England, & Fogarty, 2009). The last date partners were FTM transgender, and two men’s last study was IRB approved at every college and university sur- date partners were MTF transgender. We coded these as non- veyed and at the university where data were analyzed. heterosexual dates. Similarly, three non-heterosexually-iden- tifed women reported that their last long-term relationship was with a FTM transgender person; we counted these as Measures non-heterosexual relationships. We identified 718 same-sex hookups, including 398 The survey asked detailed information about student’s male–male hookups and 320 female–female hookups in this characteristics and attitudes as well as the students’ most dataset using the self-reported gender of the participant, and recent hookup that occurred while they were in college, the reported gender of their most recent hookup partner. How- which is the data we focus on in this study. We examined ever, approximately 7% (27 students) of male participants who variables related to students’ social activities and attitudes, reported a male partner in their last hookup also reported that prior sexual and romantic history, characteristics of, and they had vaginal sex during their last hookup or date. We cannot sentiments about student’s last same-sex hookup. Several know if these participants mistakenly entered the wrong gender variables were collected using “strongly agree,” “agree,” of their most recent hookup partner, mistakenly reported that “disagree,” and “strongly disagree” as options; for simplic- they had vaginal sex during their last encounter, were using an ity sake, we dichotomize these measures into those who alternative defnition of vaginal sex or gender, were “jokesters” “agree” and “disagree,” whether strongly or otherwise. intentionally mis-answering questions, or were thinking of Several outcomes we examined had many missing responses. two diferent hookup partners when they answered these ques- Attitude questions were not added to the survey until the fall tions. Some may identify partners or themselves as “male” of 2007, and we therefore had a smaller sample size for those or “female” instead of identifying a partner or themselves as questions than for other questions in the survey. Since this is transgender, despite the availability of “transgender” as an a select small population, to make full use of the data we did option. Upon further investigation into other questions regard- not harmonize the data according to missing outcomes on out- ing all prior sexual experience, 14 students reported past sexual come variables examined, and only deleted participants from the activity with a male partner; we retained these cases in the data based on missing responses related to key sexual identity, sample and removed the 13 who indicated they had vaginal sex partner gender, and control variables. As a sensitivity test, we with men but reported no sexual activity with men under their examined the likelihood of missing values on these variables lifetime sexual behavior. We also removed 10 students (eight for heterosexually identifed students versus non-heterosexually women and two men) missing information on race, religious identifed students and found no signifcant diferences in non- attendance, and/or mother’s education. Our fnal sample was response with only two exceptions. (Full results are available the remaining 695 students whose most recent hookup was with from authors). a same-sex partner: 383 men and 312 women.

Social Activities and Sexual Attitudes Same‑Sex Hookups and Sexual Orientation We examined several social activities and attitudes related to We identifed same-sex hookups based on the reported gen- sexuality, homosexuality, and desires for hookups that may der of the participant and their most recent hookup partner. illuminate whether student’s attitudes and characteristics We labeled sexual identity using participants’ response to align with theory related to sexual experimentation scripts, the question, “What is your sexual orientation?” They chose fraternity hazing rituals, and internalized heterosexism. We from four possible responses: “homosexual,” “heterosexual,” examined a measure of fraternity or sorority membership and “bisexual,” and “I don’t know.” Regression analyses com- agreement with the statements “Any kind of sexual activity pared those “homosexual,” “bisexual,” or “I don’t know” is ok as long as both persons freely agree to it,” “I wish there responses to “heterosexual” responses. We refer to the latter were more opportunities for hooking up at my college,” and

1 3 Archives of Sexual Behavior

“I don’t really want to be in an exclusive relationship now (whether or not you knew the person beforehand).” To exam- because I’d rather be free to date or hook up with multi- ine the prevalence of anonymous hookups related to R4, we ple people.” Social activities and attitudes related to inter- examined responses to “How well did you know the person you nalized heterosexism and prior descriptions of this group hooked up with before the day you two hooked up?” including included religious service attendance, divided into those the categories “very well,” “moderately,” “somewhat or a little who attended “never,” “1–11 times per year,” or “12 + times bit,” and “not at all.” To examine characteristics associated per year”; and agreement with the statement “My religious with performative bisexuality narratives, we explored whether beliefs have shaped and guided my sexual behavior.” Students hookups included kissing or non-groping only (including were also asked, “What is your opinion about sexual relation- “kissing” or “making out,” “you touched your partner’s breast ships between two adults of the same sex?” and “There’s or buttocks area,” or “had your breast or buttocks touched by been a lot of discussion about the way morals and attitudes your partner,” but no other sexual activity reported); whether about sex are changing in this country. If a man and a woman hookups included any type of genital contact, including have sex relationship before marriage, do you think it is;” we whether participants had oral sex, vaginal or anal sex, or hand- examined a dichotomous measure of whether participants genital stimulation (full defnitions of these terms above); and a answered “always wrong” or “almost always wrong” versus separate measure of whether hookups included vaginal or anal “wrong only sometimes” or “not wrong at all” on these two sex. We also examined whether hookups took place in public questions. We also examined whether students characterized via responses to a question “when you hooked up, where did their political views as “liberal,” “moderate” (“middle of the you go?” and whether students indicated “nowhere–we hooked road”), or “conservative.” up at a social event in plain sight” instead of one of the other options, which included “my room,” “the other person’s room,” Past Heterosexual and Same‑Sex Sexual Experience “in a private room somewhere else,” or “other.”

Past same-sex relationship and sexual experience and past het- Risk‑Taking and Sexual Assault erosexual sexual encounters may shape student’s sexual iden- tity, in alignment with developmental stages theory and theory To examine characteristics associated with internalized hetero- related to the social construction of identity. Little prior same- sexism and the college experimentation/hookup script, both of sex experience may also indicate students are engaging in sexual which indicate a high rate of substance and sexual risk-taking experimentation. We frst examined whether participant’s last during sexual encounters, we next examined risk-taking during date and participant’s last long-term relationship lasting longer same-sex hookups. We measured the number of drinks the par- than 6 months was with a non-heterosexual partner, examining ticipant had consumed before or during the hookup by totaling only those who had been on a date or formed a relationship since responses to questions about number of beers, glasses of wine, starting college. We next examined lifetime measures of partici- mixed drinks or shots, and malt beverages (Smirnof ice, Bac- pants’ experience with same-sex vaginal or anal sex (combining ardi breeze, Zima, etc.). In the latent class analysis, these were “vaginal sexual intercourse,” “anal intercourse: you penetrated divided into those who did not drink, those who drank moder- your partner” and “anal intercourse: your partner penetrated ately or those who binge drank, measured as four or more drinks you”); same-sex oral sex (combining “you performed oral sex for women and fve or more for men, using cutofs from prior on your partner” and “your partner performed oral sex on you”); research on binge drinking in hookups (Kuperberg & Padgett, and hand-genital stimulation with a same-sex partner (combin- 2017). Whether the participant used drugs was measured by the ing “you stimulated your partner’s genital with your hand” and question “What drugs did you use before or during that occa- “had your genitals stimulated by your partner’s hand”). We also sion (check all that apply)?” A response of “yes” to any of the examined whether participants ever engaged in heterosexual drugs (including marijuana, amphetamines (speed), cocaine, vaginal sex. ecstasy (x, e), heroin, mushrooms, other) resulted in a 1 or a 0 for a response of “I did not take any drugs before or during the Characteristics of Most Recent Same‑Sex Hookups hookup.” We measured whether male participants had anal sex during an encounter but answered no to the question “Did you We examined several variables related to characteristics of use a condom?” with 1 indicating both of these were true and 0 the same-sex hookup that students described in response to a indicating that they did not have anal sex during the encounter series of questions headed by the statement “For this section, or used a condom if they did. Finally, we examined whether use whatever defnition of hookup you and your friends gener- the hookup was the result of sexual assault (R5), by measuring ally use. It doesn’t have to include sex to count if you and your whether the participant answered yes to any of three statements friends would call it a hookup” and subheaded by the statement about activity during last hookup, including “Did you have “Now, some questions about the last time you hooked up with sexual intercourse that was physically forced on you?”; “Did someone you were NOT already in an exclusive relationship someone try to physically force you to have sexual intercourse,

1 3 Archives of Sexual Behavior but you got out of the situation without having intercourse?”; college degree, and current cumulative GPA, including four and “Did someone have sexual intercourse with you that you dichotomous variables: < 2.1, 2.1–3.0, 3.1–3.75, and 3.76 + . did not want when you were drunk, passed out, asleep, drugged, or otherwise incapacitated?” Analytic Strategy

Sentiments About Most Recent Same‑Sex Hookup We estimated mixed-efects logistic and linear regression models comparing heterosexually-identifed students whose Sentiments about hookups after the fact can illuminate whether last hookup was with a same-sex partner to those other-iden- the hookup was a result of sexual curiosity now satisfed after tifed students on the basis of demographic characteristics, experimentation or the beginning of some of the stages of iden- attitudes about sexuality, marriage, religion, past sexual and tity development that may lead to future changes in identity. relationship experiences, and various characteristics of stu- We examined responses to the question “looking back on this dent’s same-sex hookups. These types of models allowed hookup, how do you feel about it?” focusing on the responses us to account for clustering at the university level and were “I regret I did it” and “I’m glad I did it”; a third category “I’m estimated using the meqrlogit and mixed commands in Stata. neither glad nor regret it” was included in denominators, but not Results presented are regression-adjusted predicted means, presented separately. We also examined average responses to produced using the predict command in Stata and then gen- “How much did you enjoy the hookup overall?” with responses erating average predicted values by sexual orientation using ranging from 1 = “not at all” to 4 = “very much.” Finally, we tabstat. All models were estimated separately by gender, lim- examined dichotomized responses to whether participants indi- ited to students whose last hookup was with a same-sex part- cated some interest in response to “At the end of the hookup, ner, and controlled for age, race, mother’s education, GPA, were you interested in hooking up with this person again?” and and religious service attendance. Finally, we used the doLCA “Were you interested in having a romantic relationship with command from the Latent Class Analysis (LCA) plugin in the person you hooked up with after you hooked up?” with the Stata (Lanza, Dziak, Huang, Wagner, & Collins, 2015; LCA responses “Yes, I was defnitely interested,” and “Maybe, it had Stata Plugin, 2015) to conduct a LCA to determine whether some appeal” counted as a “1,” and “Possibly, I didn’t really certain characteristics correlated with underlying typologies know yet,” and “No, I wasn’t at all interested” counted as 0. defning distinct groups among heterosexually-identifed stu- dents who engaged in same-sex hookups. LCA methods can Demographic Variables illuminate latent typologies or “classes” and are preferable over more crude but analogous factor analysis or cluster anal- We present results related to selection into a heterosexual ysis methods in ofering analyses that are more in line with identity among students whose last hookup was same-sex what is theoretically meaningful in social science research by demographic characteristics and later controlled for these (Hagenaars & Halman, 1989). Classes are assumed to be characteristics in subsequent regression analyses. Related to categorical, unlike factor analysis which assumes underlying development stages described in R2, age was examined in “factors” are continuous (Collins & Lanza, 2010). Results four groups: 18–19, 20–21, 22–23, and 24 + , although the presented are the predicted probability that a participant was Latent Class Analysis (discussed further below) examined a member of a specifc class; these are not interpreted simi- only whether participants were 18 or an older age. Related larly to factor loadings. Rather, values close to 1 or close to to R4, race was examined in response to the question “If you 0 indicate a strong relationship between a given variable and had to pick one racial or ethnic group to describe yourself, the latent class, but the distribution of probabilities across which would it be?” with results separated into White, Black, classes must also be examined to determine which variables Hispanic (including original categories Mexican–American are signifcant (Collins & Lanza, 2010). and Other Hispanic), Asian [including original categories Chinese (from USA, PRC, Taiwan, Hong Kong, Singapore, etc.), Japanese, Korean, Filipino, Vietnamese, South Asian Results (Indian, Pakistani, etc.), Other Asian/Pacifc Islander], and Other race (including original categories Native American Table 1 shows the self-identifed sexual orientations or “sexual Indian/Native Alaskan and Other). To account for past dif- identity” of participants who engaged in same-sex hookups. ferences in hooking up and risk-taking found by past research Of the 383 male–male hookups examined, 12% (or 45) were (Kuperberg & Padgett, 2015, 2016, 2017), we also examined embarked on by a heterosexually-identifed male participant, and controlled for participant’s mother’s highest level of edu- and among the 312 female–female hookups, 25% (or 77) were cation by including a dichotomous measure of whether their undertaken by a heterosexually-identifed female participant. mother had a bachelor’s degree or graduate degree versus no In the broader sample, same-sex hookups comprised 8.4% of

1 3 Archives of Sexual Behavior the 4746 most recent hookup experiences reported by men, and Characteristics of Students Who Hookup 3.3% of the last 9884 most recent hookup experiences reported with Same‑Sex Partners, by Sexual Identity by women. In general, women who hooked up with women had more variation in their sexual identities compared to men who Demographic Variables hooked up with men. Sixty-eight percent of men who hooked up with men identifed as homosexual, versus only around 39% Table 2 presents odds ratios from mixed-effects models of women who hooked up with women. Around twice as many predicting a heterosexual identity among students’ whose women as men identifed as heterosexual (25% of women vs. most recent hookup was with a same-sex partner, provid- 12% of men) or bisexual (29% of women vs. 13% of men). ing evidence of demographic selection into a heterosexual Rates at which the participant was unsure of their sexuality were identity, with distinct patterns for men and women. Heter- approximately equal, at around 7% for both men and women. osexually-identifed women were signifcantly younger, but men showed no signifcant diference by age in a hetero- sexual identity. Women had no signifcant diference in a heterosexual identity by race, but Asian men were less likely to identify as heterosexual than White men. Class impacted women’s sexual identities, but not men’s; women with a col- Table 1 Self-identifed sexual orientation of students whose last hookup was same sex lege-educated mother were signifcantly less likely to identify as heterosexual. GPA was not related to identity. Men Women N % N % Social Activities and Attitudes

Homosexual 262 68.4 122 39.1 Religious, sexual, and political attitudes were correlated with Bisexual 51 13.3 90 28.9 sexual identity among students hooking up with same-sex I’m not sure 25 6.5 23 7.4 partners. Table 2 demonstrates that religious service attend- Heterosexual 45 11.8 77 24.7 ance was positively and signifcantly correlated with a hetero- Total N 383 312 sexual identity among both men and women. Table 3 presents

Table 2 Mixed-efects logistic Men Women regressions predicting whether (N = 383) (N = 312) students who engaged in same-sex hookups identifed as Odds ratios 95% Conf- Odds ratios 95% heterosexual (1) or homosexual, dence interval Confdence bisexual, or unsure (0) interval

Age 18–19 (ref) Age 20–21 1.44 0.63–3.31 0.54 0.29–1.01 Age 22–23 1.42 0.51–3.93 0.24** 0.09–0.66 Age 24+ 2.23 0.74–6.74 0.25* 0.09–0.74 White (ref) Black 0.25 0.05–1.32 0.80 0.27–2.34 Hispanic 0.24 0.05–1.11 1.33 0.57–3.11 Asian 0.12* 0.01–0.98 1.78 0.67–4.70 Other race 0.67 0.12–3.80 0.00 0.00–0.00 Mother BA+ 1.14 0.56–2.30 0.51* 0.28–0.92 GPA < 2.1 2.33 0.27–18.44 0.95 0.23–3.82 GPA 2.1–3.0 4.39 0.88–21.98 1.36 0.46–4.03 GPA 3.1–3.75 3.40 0.72–16.09 0.77 0.27–2.23 GPA 3.75 + (ref) Attends religious services sometimes 2.15* 1.05–4.41 1.97* 1.07–3.63 Attends religious services 1 +/month 4.62* 1.11–19.14 3.22* 1.28–8.10

*p < .05; **p < .01; ***p < .001

1 3 Archives of Sexual Behavior signifcant diferences in social activities and attitudes, with or relationship was with a non-heterosexual partner, or that results presented being predicted percentages estimated from they had past experience of same-sex hand to genital stimu- mixed-effects logistic regressions. Fraternity and sorority lation, oral sex, and vaginal or anal sex. They were also sig- membership did not signifcantly difer by sexual identity nor nifcantly more likely to report past heterosexual vaginal sex. did attitudes related to whether participants believed any con- sensual sex was ok, wanted more opportunities to hookup on Characteristics of the Hookup and Later Sentiments campus, or wanted to avoid relationships so that they could date and hookup with multiple people. Four attitudes signif- Table 4 shows prior knowledge of partners, sexual activity, cantly diferentiated heterosexually-identifed men from non- assault, and substance use during encounters, sentiments heterosexually-identifed men: Heterosexuals were more likely about the encounter, and diferences by sexual identity among to agree that “religion informs my sexual decisions”; less likely students whose last hookup was with a same-sex partner. to agree same-sex relationships were never wrong; more likely Men with a heterosexual identity knew their same-sex part- to agree that premarital sex was always wrong; and more likely ners signifcantly better than men with a non-heterosexual to hold conservative political views. For female participants the identity, but women had no signifcant diferences in prior same patterns held true, except that women additionally were knowledge of partner. Almost a third of same-sex hookups more likely to identify as liberal if they had a non-heterosexual among heterosexually-identifed women took place in public, identity. signifcantly more than among non-heterosexually-identifed women, but this rate did not signifcantly difer among men. Past Non‑heterosexual and Heterosexual Sexual Experience Both men and women with a heterosexual identity were sig- nifcantly more likely to only engage in lower-order sexual All measures related to sexual and relationship experience activity during encounters and less likely to engage in geni- presented in Table 3 were signifcantly related to students’ tal contact, although there were no signifcant diferences in sexual identity. Among both men and women whose last prevalence of anal or vaginal sex, unprotected anal sex, or hookup partner was same-sex, those with a heterosexual sexual assault. identity were signifcantly less likely to report their last date

Table 3 Characteristics of students who hookup with same-sex partners, by gender and sexual identity (regression-adjusted means predicted from mixed-efects logistic regressions) Men Women Homosexual/ N Heterosexual N Homosexual/ N Heterosexual N bisexual/ bisexual/ unsure unsure

Sorority or fraternity member 7.7 337 17.8 44 4.1 235 4.2 77 Agrees any consensual sex ok 91.1 335 91.7 44 89.2 235 95.2 74 Wants more opportunities to hook up on campus 51.4 333 43.1 43 28.3 233 29.0 74 Doesn’t want to be in an exclusive relationship so can 34.3 332 40.5 44 27.1 235 35.5 75 date/hookup with multiple people Agrees religion shapes my sexual decisions 16.3* 334 34.1 44 16.4* 235 31.4 75 Agrees Same-sex sexual relations are not wrong 97.8*** 304 73.3 43 97.2** 220 85.4 68 Agrees premarital sex is wrong 2.0** 302 14.7 42 2.9* 214 8.5 65 Liberal political views 77.0 311 58.6 43 88.2** 225 66.3 71 Conservative political views 3.9** 311 24.1 43 1.6* 225 9.4 71 Last date was with non-heterosexual partner (if dated) 91.3*** 265 54.6 34 80.6*** 187 16.3 59 Last long-term relationship was with non-heterosexual 90.1* 102 69.9 24 68.8*** 141 12.5 32 partner (if long-term relationship) Past experience same-sex hand to genital stimulation 98.7*** 338 51.5 45 93.1*** 235 22.7 45 Past experience same-sex oral sex 97.6*** 338 54.0 45 82.0*** 235 21.9 45 Past experience same-sex vaginal or anal sex 80.3*** 338 17.3 45 43.1*** 235 4.1 77 Past heterosexual vaginal sex 21.8* 338 33.7 45 57.3** 235 75.4 77

Regression adjusted to standardize for age (Ref: 18–19), race (Ref: White), religious attendance (Ref: Never), GPA (Ref: 3.75 +) and mother has a college degree (Ref: Does not) *p < .05; **p < .01; ***p < .001

1 3 Archives of Sexual Behavior

Table 4 Characteristics of last same-sex hookup partner and last same-sex hookup, by gender and sexual identity (regression-adjusted means predicted from mixed-efects logistic regressions) Men Women Homosexual/ N Heterosexual N Homosexual/ N Heterosexual N bisexual/ bisexual/ unsure unsure

Knew partner 338 45 235 77 Very well 12.2 17.6 33.0 36.8 Moderately well 13.1** 32.0 29.3 18.8 Somewhat or a little bit 45.5 40.7 29.4 33.6 Not at all 29.5* 10.9 8.5 10.7 Hookup took place in public 5.9 336 9.7 45 11.2*** 233 29.4 76 Hookup included kissing or above-the-waist groping only 6.7*** 336 29.0 42 28.1** 232 49.2 76 Hookup included any genital contact 93.3*** 338 71.0 45 71.9** 235 50.8 77 Hookup included vaginal or anal sex 37.7 338 24.7 45 17.4 235 26.0 77 Hookup included anal sex without a condom 6.4 338 10.5 45 – – – – Rape (forced, incapacitated or attempted) 5.7 337 12.1 42 2.3 234 3.4 75 # Of drinks before or during hookup 2.4** 338 4.3 45 3.0 235 3.9 77 Used drugs during or before hookup 15.6 338 18.0 45 11.4 235 6.9 77 Regrets hookup 13.8 260 13.3 35 9.8 172 13.8 63 Glad about hookup 38.0* 260 18.8 35 61.5* 172 43.7 63 How much Enjoyed Hookup overall (1 = not at all, 4 = very 2.1* 317 1.9 40 2.5** 226 2.2 69 much) Interested in hooking up again 54.2 239 43.4 39 74.1*** 216 46.9 68 Interested in relationship with hookup partner 34.9 335 30.3 45 50.8** 235 35.9 77

Regression adjusted to standardize for age (Ref: 18–19), race (Ref: White), religious attendance (Ref: Never), GPA (Ref: 3.75 +) and mother has a college degree (Ref: Does not) *p < .05; **p < .01; ***p < .001

In terms of other types of risk-taking, men who identi- heterosexually-identifed students hooking up with same- fed as heterosexual had consumed a signifcantly higher sex partners. Second, we selected variables that the above number of drinks before or during hookup, but women did analyses indicated were unusually common among hetero- not, and no signifcant diferences in drug use occurred by sexual-identifying students who hooked up with same-sex sexual identity. There were no signifcant diferences among students, compared to non-heterosexually-identifed students men and women by sexual identity in terms of regret about who hooked up with same-sex students. the hookup, but both men and women with a heterosexual Table 5 shows ft statistics for the models with diferent identity were signifcantly less likely than those with another number of classes of heterosexuals who hooked up with identity to be glad about the hookup or to describe it as enjoy- same-sex partners. AIC and adjusted BIC values gave support able, and women with a heterosexual identity were also sig- for a 6-class solution; we gave greater weight to adjusted BIC nifcantly less likely to be interested in a repeat hookup or a as recommended by Tein, Coxe, and Cham (2013). Table 6 relationship with the same-sex partner from their last hookup. shows the variable correlations with each class. Numbers represent the probability of certain characteristics occurring Latent Class Analysis among members of that “class”; probabilities close to 0 or 1, or unusually high for that variable, are especially of note To investigate whether heterosexually-identifed students (Collins & Lanza, 2010) and we bold probabilities that are who hookup with same-sex partners can be described as unusually high. Rape or attempted rape during a hookup and comprising distinct types, we conducted a latent class anal- fraternity/sorority membership was not strongly correlated ysis of these students (N = 122) using several of the vari- with any of the classes. ables described above. This analysis did not include non- The frst three classes, which we describe as “Experimenta- heterosexually-identifed students. We selected variables tion/Early Stages,” shared the fact that they were mostly private for the latent class analysis based on several criteria: frst, encounters that took place among those who agreed premarital several variables aligned with theoretical explanations for sex, consensual sex, and homosexuality were not almost always

1 3 Archives of Sexual Behavior

Table 5 Model ft statistics for Number of Log-likelihood AIC BIC Adjusted BIC Entropy R-Sqd the optimal number of classes classes of heterosexuals hooking up with same-sex partners 1 − 1328.99 1549.28 1616.58 1540.69 NA 2 − 1288.44 1518.19 1655.59 1500.66 .77 3 − 1240.24 1471.78 1679.28 1445.31 .80 4 − 1207.33 1455.97 1733.57 1420.56 .90 5 − 1185.14 1461.59 1809.29 1417.23 .90 6 − 1157.79 1456.88 1874.68 1403.57 .91 7 − 1140.92 1473.15 1961.05 1410.90 .93 8 − 1123.62 1488.56 2046.56 1417.36 .93

Bold values are unusually high proportions or always wrong, were not particularly religious, and whose a future relationship. This group was most likely to describe actions may be said to conform with a sexual experimentation themselves as politically middle of the road, to have been drink- script, or may suggest earlier stages of non-heterosexual identity ing moderately, and had the least overall support for any consen- development. Comprising 29% of heterosexually-identifed stu- sual sex being ok of any class, although 76% still agreed it was dents who hooked up with same-sex partners, the frst and larg- ok, and all were in support of premarital sex. Like the second est class, which we refer to as “wanting more,” were those who class, some said religion infuenced their sexual decisions, and very much enjoyed the encounter, had the second highest rate not all agreed homosexual relationships were always ok. Apart of wanting a later relationship with the partner (57%), and were from their low level of enjoyment and level of inebriation, what the most likely to have engaged in prior same-sex penetrative distinguished this class from the frst two was their low level of vaginal or anal sex; although only 30% had previously engaged sexual activity; 82% did not proceed beyond kissing and grop- in this activity; this was the highest correlation for any class. A ing during the encounter. All knew their partners moderately total of 68% had some kind of genital contact with their partner or very well before they hooked up with them. Comprising 9% during the encounter. 42% had been binge drinking but nearly of heterosexually-identifed participants who hooked up with half did not drink during the encounter; the second highest rate same-sex partners, this was one of the smaller classes. for any class. Comprising 21% of heterosexually-identifed participants The second largest group, comprising 22% of partici- who hooked up with same-sex partners, the fourth class, whom pants, whom we describe as “drunk and curious,” consisted we refer to as “maybe for show,” conformed closely to theory of those with little prior homosexual sexual experience (5%) regarding performative bisexuality. All participants in this and who were especially likely to be binge drinking during class were women, 70% were age 18, and 98% of these encoun- the encounter (72%). They had the highest rate of describing ters took place in public “at a social event in plain sight.” Stu- themselves as politically liberal among the frst three classes, dents in this class were also most likely to be binge drinking and the second highest overall, and were strong supporters (84%), did not have any prior experience with same-sex vagi- of premarital sex and consensual sex generally, with 96% nal sex (0%), mostly only kissed or groped breasts or buttocks agreeing both were ok, although 38% admitted to religious during the encounter (91%), and only a minority were inter- infuence on their sexuality, and 20% thought homosexuality ested in a future relationship with their hookup afterward (9%) was almost always or always wrong. While this group had the although 31% enjoyed the encounter “very much.” Students highest rates of engaging in genital contact at 80%, unlike the in this class universally described themselves as politically frst class, they mostly did not want a future relationship with liberal, and agreed premarital sex, any consensual sex, and their last same-sex hookup partner, with only 4% wanting homosexual relations were ok. They were the least religious such an encounter, and most commonly said they enjoyed of any class, with 82% never attending religious services, and the hookup “somewhat” (57%), while 23% enjoyed it “not at only 8% stating religion informed their sexual decisions. all” and none said they enjoyed it “very much.” By contrast, The last two classes comprised those who were religious, in the frst group, over half enjoyed the hookup “very much” and whose religious identity perhaps conficted with a non- and almost all the remainder enjoyed it “somewhat.” This heterosexual identity. The ffth group we term “loved it, but group was most likely to know their partner “not at all” before religious.” This group consisted of mostly women (92%), that night (34%). who especially enjoyed their encounter compared to other The third group, which we describe as “little enjoyment,” classes, but with 45% attending services at least once a month was least likely to report enjoying the encounter with 81% and the remaining 55% attending at least once a year but less stating they enjoyed it “very little” and almost none wanting than once a month, they were also the most likely to attend

1 3 Archives of Sexual Behavior

Table 6 Latent class analysis: Heterosexual students who hookup with same-sex partners Experimentation/early stages College scripts Religious Wanting more Drunk and Little enjoyment Maybe for show Loved it, but Just not curious religious who I can be

Female 0.36 0.75 0.49 1.00 0.92 0.02 Age 18 0.35 0.20 0.26 0.70 0.65 0.25 Fraternity or sorority member 0.22 0.05 0.00 0.00 0.00 0.14 Religious services: never 0.29 0.48 0.33 0.82 0.00 0.01 Religious services 1–11 ×/year 0.65 0.44 0.66 0.11 0.55 0.98 Religious services 1 +/month 0.06 0.08 0.00 0.07 0.45 0.01 Religion informs sexual views 0.06 0.38 0.42 0.08 0.57 0.87 Premarital sex wrong 0.00 0.04 0.00 0.00 0.27 0.70 Homosexual relations wrong 0.00 0.20 0.17 0.00 0.33 0.98 Any consensual sex is OK 1.00 0.96 0.76 1.00 0.92 0.88 Politically liberal 0.67 0.84 0.39 1.00 0.35 0.42 Politically moderate 0.18 0.00 0.61 0.00 0.44 0.15 Politically conservative 0.14 0.16 0.00 0.00 0.21 0.43 At end of hookup interested in relationship 0.57 0.04 0.01 0.09 0.71 0.13 How much enjoyed hookup overall None 0.05 0.23 0.00 0.07 0.00 0.14 Very little 0.00 0.20 0.81 0.14 0.00 0.41 Somewhat 0.44 0.57 0.18 0.48 0.22 0.44 Very much 0.51 0.00 0.01 0.31 0.78 0.01 Did not drink 0.49 0.16 0.27 0.07 0.51 0.25 Moderate drinking 0.09 0.13 0.72 0.09 0.07 0.25 Binge drinking 0.42 0.72 0.01 0.84 0.42 0.50 Rape or attempted rape during hookup 0.08 0.16 0.00 0.00 0.00 0.16 Prior experience same-sex vaginal/anal sex 0.30 0.05 0.00 0.00 0.00 0.00 Only kissed or groped breasts or buttocks 0.32 0.20 0.82 0.91 0.35 0.43 Took place “at a social event in plain sight” 0.00 0.01 0.26 0.98 0.33 0.24 Knew partner before hookup Not at all 0.00 0.34 0.00 0.20 0.00 0.25 Somewhat or a little bit 0.10 0.25 0.00 0.14 0.16 0.48 Moderately 0.67 0.15 0.83 0.29 0.39 0.26 Very well 0.23 0.26 0.17 0.36 0.45 0.01 Proportion in class 0.29 0.22 0.09 0.12 0.21 0.07

Bold values are unusually high proportions religious services regularly. Further, 57% stated that religion less prior same-sex sexual experience (0%), and higher rate informed their sexual views. Regarding sexuality, their views of religiosity. were more mixed: A signifcant minority thought homosexu- The fnal and smallest class (7%), which we refer to “just ality (33%) and premarital sex (27%) was almost always or not who I can be,” comprised those whose characteristics cor- always wrong, but 92% also stated any consensual sex was ok. responded with theory related to internalized heterosexism. About two-thirds of this class were age 18 (65%). This class Almost all men (98%), this class was not likely to attend reli- distinguished itself by being most interested in a relation- gious services at least monthly like the prior class, but 98% ship after a hookup (71%), and the most likely to state they attended services between 1 and 11 times in the past year, and enjoyed the hookup very much (78%), while also the most they had the highest rate of stating religion informed their sexual likely to have not been drinking during the hookup (51%). views (87%). This group was also almost universally likely to This group shared much in common with the frst class, but state homosexual relations were almost always or always wrong was distinguished from the frst class by their younger age, (98%), and 70% of this group thought premarital sex was almost

1 3 Archives of Sexual Behavior always or always wrong. This group was most likely to describe Sexual Experimentation Scripts themselves as politically conservative (43%). Although only some wanted a relationship with their same-sex partner after Our frst research question asked whether fndings aligned with the hookup (13%), this group mostly enjoyed the hookups some- a sexual experimentation script, which could include performa- what or very little, distinguishing themselves from the ffth class tive bisexuality or fraternity hazing rituals. We found little sup- who were more likely to say they enjoyed it very much. port for fraternity hazing rituals being a factor; heterosexually- identifed students were not more likely to be in fraternities or sororities. Students may not conceptualize these encounters Discussion as “hookups.” However, many findings supported a sexual experimentation/partying narrative. For instance, heterosex- This study was the frst to conduct a systematic comparison ually-identifed men drank more during encounters. Hetero- and analyses of heterosexually-identifed students who hook sexually-identifed students reported enjoying encounters less up with same-sex partners, and the specific circumstances but were not more likely to report regretting the experience, under which those hookups occur. The large, rich dataset we perhaps indicating some experimental nature of many of these draw upon is superior to other samples that relied on snowball encounters; students did not regret experimenting, but some sampling or recruitment through specifc LGB venues, because found that they did not enjoy that experiment, reafrming a the size allowed for us to examine rare groups, including het- heterosexual identity. Heterosexually-identifed women were erosexually-identifed students who hookup with same-sex less likely to be interested in a repeat performance or future partners, and features of those encounters. However, it was not relationship with their last same-sex hookup partner, also sup- without limitations. The dataset is not representative of col- porting an experimentation narrative. Some fndings challenged lege students in general. Within colleges, the inclusion of some a sexual experimentation narrative, as sexual identity did not courses addressing gender and sexuality in the sample likely led correlate with belief in the acceptability of any consensual sex, to greater selection among students who were questioning their wanting more hookup opportunities or wanting to avoid exclu- sexual orientation or generally more interested in sexual topics sive relationships to hookup with multiple people. Performative than other students, leading to some skewing of results, espe- bisexuality explanations were also supported by our fndings; cially overall rates at which sexual identities may occur. The heterosexually identifed women were more likely to hookup survey only included college students, and only asked about one with same-sex partners in a public space, consistent with prior hookup they experienced, and cannot tell us about same-sex descriptions. hookups among self-identifed heterosexuals who are not in college, or the trajectory of sexual identity formation. Finally, Sexual Identity Development and Internalized the large number of tests we conducted may have increased Heterosexism Type-1 errors (false positives) while the small sample may have increased Type-2 errors (false negatives). Our second research question asked whether student’s charac- Heterosexually-identified students who hookup with teristics aligned with sexual identity theory, either that related same-sex partners comprised a substantial number of same- to sexual identity development, or the social constructionist sex hookups. In these data, heterosexually-identifed students model. When compared with those other-identifed students, accounted for approximately one in nine participants of the heterosexually-identifed women engaging in same-sex hookups most recent same-sex hookups among college men, and one tended to be younger and were less likely to have had prior non- in four of the most recent same-sex hookups among college heterosexual sexual, dating, or relationship experience, while women. Findings suggest that survey questions designed to more likely to have had heterosexual vaginal sex. These fndings capture sexual orientation data may be instead measuring support both theories related to sexual identity development, sexual identity. Unlike sexual orientation, which describes and the social construction of sexuality, which indicate sexual sexual feelings toward one or more genders and may be bio- identity is developed and reafrmed via experiences with same- logically based, sexual identity is adopted by individuals sex and other-sex partners. within a specifc sociohistorical context, precluding some Our third research question focused on internalized hetero- with same-sex attractions from taking on such an iden- sexism. Internalized heterosexism was refected in the lower tity. Same-sex sexual behavior may also be undertaken by acceptance of same-sex sexual relations among those with a those who may not have an attraction to same-sex partners heterosexual identity and, in line with literature on heterosex- in specifc sociohistorical contexts where sexual “scripts” ism, religious service attendance and religious infuence on encourage same-sex sexual contact and sexual experimenta- sexual behavior also correlated with a heterosexual identity tion (Simon & Gagnon, 1986, 2003). The script of women for both men and women, as did conservative political and engaging in same-sex low-level hookups at public parties in sexual views. In contrast to theory on internalized heterosex- college is one such social script (Sanday, 2007). ism and risk-taking, heterosexual students were not more

1 3 Archives of Sexual Behavior likely to have unprotected sex or use drugs during encounters, partner, prior homosexual experience, and drinking during but men (who we found in the latent class analysis comprised the encounter. While the frst class (“wanting more”) wanted almost all of those who could be described as experiencing relationships with their partners and may later change their internalized heterosexism) drank more alcohol. identity, the second class (“drunk and curious”) seems to have experimented perhaps for the sake of sexual experi- Race and Knowledge of Partners, Sexual Assault mentation, and while they enjoyed their experiment, did not desire a relationship with their hookup. In the future, they Our data also allowed us to systematically examine racial dif- may retain a heterosexual identity or change it in reaction to ferences in the same-sex hookups of heterosexually-identifed a same-sex hookup or relationship they feel more strongly college men and how well they knew their partner. Our fourth about pursuing beyond a sexual encounter. These frst two research question centered on whether race and knowledge of classes may be akin to Cass’ (1996) “special case” pathway partners aligns with research of men who have sex with men for those identifying as heterosexual in early stages of iden- anonymously on the “down low.” In contrast with a media tity development, where students view themselves as hetero- and academic research focus on African-American men who sexual apart from this one partner or single event but may have sex with men (Ward, 2015), we found that White men later form a relationship that leads to adoption of an LGB are signifcantly more likely than Asian men and no diferent sexual identity. The third class (“little enjoyment”) consisted than Black men to report a “heterosexual” identity despite of those who experimented and perhaps confrmed a hetero- engaging in same-sex hookups. Other fndings also call into sexual identity after not enjoying the encounter and ended the question the sample selection methods of researchers seeking encounter before proceeding to higher order sexual activity. to examine men on “the down low”; heterosexually-identi- This class may chalk this up as an experience to check of fed men knew their partners better than those with other their college experimentation list and retain a heterosexual identities, indicating these were not necessarily the random identity, or instead may later engage in more enjoyable same- hookups described in past studies that relied on, for instance, sex encounters that lead to a shift in identity. samples of men who hooked up with men in public places. The fourth class (“maybe for show”) was women, often 18, Research has found that students who knew their partners bet- who kissed and “made out” with same-sex partners in public ter tended to take more risks such as unprotected sex (Kuper- settings. These students may be engaging in performative berg & Padgett, 2017), indicating the importance of studying hookups (Hamilton, 2007) in accordance with performative encounters among heterosexually identifed men who have bisexuality social scripts aimed at attracting men, but may sex with men they know well, which may not be captured via also be using these opportunities to experiment with same-sex more commonly used methods of recruitment that focus on attraction (Ward, 2015); about one in 10 stated they wanted a anonymous hookups. Our ffth research question also asked future relationship with the partner, and a third enjoyed the whether these encounters were a result of sexual assault, but encounter “very much.” Finally, the last two classes were we do not fnd support for a higher rate of sexual assault religious students whose strong religious participation or among those with a heterosexual identity. infuence on their behavior likely afected sexual identity. Comprising over one-fourth of students who identifed as Who Comprises Heterosexual Students Who Hookup heterosexual, the majority (“loved it, but religious”) did not with Same‑Sex Partners? conform to theory related to internalized heterosexism, but instead may have been conflicted about taking on a non- Our final research question asked whether students who heterosexual identity given their frequent religious attend- reported hooking up with same-sex partners, but identifed as ance. They were also especially young and inexperienced heterosexual, could be divided into certain “types” described in same-sex relationships and had a high level of enjoying by the above theories, and we found through latent class the encounter; this group also conforms to theory on sexual analyses that they comprised six groups. The frst three were identity development and may take on a non-heterosexual those that can be classifed as various types of mostly pri- identity later in life, but may delay that stage compared to vate experimentation, which may be correlated with engaging less religious students due to their religious engagement. in a college experimentation script, and/or the early stages The fnal class (“just not who I can be”), the smallest class, of identity development and experiences that can lead to a which was almost all male, were those who had strong views later identity change. Comprising 60% of same-sex hookups against homosexuality, which has been termed “internalized among heterosexuals, these hookups did not take place in heterosexism.” public for the most part and took place among students with Past theory on sexual scripts, identity development theory positive views of premarital sex and homosexuality. The (whether via stages or socially constructed), and internal- three classes difered on the basis of students’ later senti- ized heterosexism all contribute to the patterns we fnd but ments about the encounter, desire for a relationship with that describes distinct groups who together comprise those who

1 3 Archives of Sexual Behavior identify as heterosexual but hookup with same-sex part- (2016) found that among a random sample of young adults ners. Past research has usually been qualitative and gener- aged 13–20, who have grown up in a world where legal gay ally focused on only one of these groups and/or theory of a unions have been available in some U.S. states for as long as heterosexual identity among those who engage in same-sex they can remember (civil unions became legal in Vermont in behavior, such as women hooking up with women at fra- 2000; same-sex marriage in Massachusetts in 2004), only 48% ternity parties (Hamilton, 2007). Our data allowed a more identify as “completely heterosexual,” down from 65% among comprehensive quantitative analysis which permitted us to those aged 21–34 (Laughlin, 2016). As young adults come reveal the degree to which these somewhat diferent theories of age in a new social system where gay marriage is legal, may describe college students accurately, and the prevalence and as progress continues to be made in anti-discrimination of certain “types” among students. legislation, patterns related to sexual identity will likely con- tinue to shift and should be the subject of continuing research Implications among college students and in other settings and populations. We suggest longitudinal studies to examine the evolution of The degree to which students and individuals choose to adopt sexual identity, meaning-making, and heterosexually identi- a non-heterosexual identity varies by social context and the fed adults engaging in same-sex hookups. circumstances of the encounter, with college being a context particularly fraught with specifc sexual scripts institutional- Acknowledgements The authors thank Joseph Padgett, R. James Leis- ter, Lonnie McLendon, Rachel Madsen, and Stephanie Pruitt for their ized into the fabric of social life (Ward, 2015). We found most research assistance on this project, as well as Sara Crawley and Joseph students who engaged in same-sex hookups but identifed Padgett for their helpful advice. This project was funded by a Univer- as heterosexual could be described as privately experiment- sity of North Carolina at Greensboro New Faculty Research Grant and ing and/or having religious conficts with assuming an LGB a New Faculty Summer Excellence Award Grant, as administered by the Ofce of Sponsored Programs. An earlier version of this paper was identity. Some of these students may later change their sexual presented at the Southern Sociological Society Conference, Charlotte, identity, but others will retain a heterosexual identity. Theory NC, 2014. related to performative bisexuality and internalized hetero- sexism described only a minority of these students (12 and 7%, respectively), while research on Black men on the “down low” and men having anonymous hookups with men exclude References most same-sex hookups between those who identify as het- erosexual. Research limited to those self-identifying as LGB Amadio, D. M. (2006). Internalized heterosexism, alcohol use, and will also miss a signifcant minority of those who engage in alcohol-related problems among lesbians and gay men. Addictive Behaviors, 31, 1153–1162. same-sex sexual encounters. Findings also suggest that not Armstrong, E. A., England, P., & Fogarty, A. C. K. (2009). Orgasm in everybody who engages in same-sex behavior is “secretly college hook ups and relationships. In B. Risman (Ed.), Families gay,” but rather may be engaging in socially scripted sexual as they really are (pp. 362–377). New York: Norton. experimentation that will not have long-term implications for Baumeister, R. F. (2000). Gender diferences in erotic plasticity: The female sex drive as socially fexible and responsive. Psychological their identity, while others may retain a heterosexual iden- Bulletin, 126, 347–374. tity (rather than a LGB identity) to resolve conficts between Baunach, D. M. (2012). Changing same-sex marriage attitudes in their religious beliefs and community standards, and same- America from 1988 through 2010. Public Opinion Quarterly, 76, sex sexual desires and actions. For a minority, a heterosexual 364–378. Bleich, S., & Taylor-Clark, K. (2005). Black men on the “down-low” identity seems related to internalized heterosexism, and for and the HIV epidemic: The need for research and intervention men, incongruence in identity and behavior is also associated strategies. Harvard Journal of African American Public Policy, with higher alcohol use during same-sex encounters. These 11, 13–20. fndings have clinical implications for those experiencing Blumstein, P. W., & Schwartz, P. (1977). Bisexuality: Some social psy- chological issues. Journal of Social Issues, 33, 30–45. distress regarding same-sex hookups and/or sexual identity. Bogle, K. A. (2008). Hooking up: Sex, dating, and relationships on In recent years as LGB identities have become more campus. New York: New York University Press. socially acceptable, the types of people selecting into a non- Bower, J., Gurevich, M., & Mathieson, C. M. (2002). (Con)tested iden- heterosexual identity may have changed, as these identities tities: Bisexual women reorient sexuality. Journal of Bisexuality, 2(2–3), 23–52. have been publicly adopted by a wider swath of the population Boykin, K. (2005). Beyond the down low: Sex, lies, and denial in Black (Baunach, 2012; Laughlin, 2016). This dataset was collected America. New York: Avalon. during a certain historical-time period (2005–2011), during Brady, S., & Busse, W. J. (1994). The Gay Identity Questionnaire: A which gay rights activity was both energized and the subject of brief measure of homosexual identity formation. Journal of Homo- sexuality, 26(4), 1–22. polarizing debate as gay marriage was being adopted by more Budnick, J. (2016). “Straight girls kissing”? Understanding same-gen- and more states, but not yet legal nationwide. A research report der sexuality beyond the elite college campus. Gender & Society, collected after the legalization of gay marriage nationwide 30, 745–768.

1 3 Archives of Sexual Behavior

Cass, V. C. (1979). Homosexuality identity formation: A theoretical seeking, and substance use. The Counseling Psychologist, 36, model. Journal of Homosexuality, 4, 219–235. 595–614. Cass, V. C. (1996). Sexual orientation identity formation: A western Kaufman, J. M., & Johnson, C. (2004). Stigmatized individuals and the phenomenon. In R. P. Cabaj & T. S. Stein (Eds.), Textbook of process of identity. Sociological Quarterly, 45, 807–833. homosexuality and mental health (pp. 239–266). Washington, DC: Kimmel, M. S. (2008). Guyland: The perilous world where boys become American Psychiatric Press. men. New York: Harper. Centers for Disease Control and Prevention. (2010). HIV among African Kimmel, M., & Plante, R. F. (2002). The gender of desire: The sexual American gay and bisexual men. Retrieved from http://www.cdc. fantasies of women and men. Advances in Gender Research, 6, gov/hiv/risk/racia​lethn​ic/bmsm/facts​/index​.html. 55–77. Chandra, A., Copen, C. E., & Mosher, W. D. (2013). Sexual behavior, King, J. L. (2004). On the down low: A journey into the lives of “straight” sexual attraction, and sexual identity in the United States: Data black men who sleep with men. New York: Broadway Books. from the 2006–2010 National Survey of Family Growth. In A. Koblin, B. A., Torian, L. V., Guilin, V., Ren, L., MacKellar, D. A., & K. Baumle (Ed.), International handbook on the demography of Valleroy, L. A. (2000). High prevalence of HIV infection among sexuality (pp. 45–66). Dordrecht: Springer. young gay men who have sex with men in New York City. AIDS, Cole, S. W., Kemeny, M. E., Taylor, S. E., & Visscher, B. R. (1996). 14, 1793–1800. Elevated physical health risk among gay men who conceal their Kuperberg, A., & Padgett, J. E. (2015). Dating and hooking up in col- homosexual identity. Health Psychology, 15, 243–251. lege: Meeting contexts, sex, and variation by gender, partner’s Collins, L. M., & Lanza, S. T. (2010). Latent class and latent transition gender, and class standing. Journal of Sex Research, 52, 517–531. analysis: With applications in the social, behavioral, and health Kuperberg, A., & Padgett, J. E. (2016). The role of culture in explain- sciences. Hoboken, NJ: Wiley. ing college students’ selection into hookups, dates, and long-term Diamond, L. M. (2003). Was it a phase? Young women’s relinquish- romantic relationships. Journal of Social and Personal Relation- ment of lesbian/bisexual identities over a 5-year period. Journal ships, 33, 1070–1096. of Personality and Social Psychology, 84, 352–364. Kuperberg, A., & Padgett, J. E. (2017). Partner meeting contexts Diamond, L. M. (2005). “I’m straight, but I kissed a girl”: The trouble and risky behavior in college students’ other-sex and same-sex with American media representations of female-female sexuality. hookups. Journal of Sex Research, 54, 52–55. Feminism and Psychology, 15, 104–110. Lanza, S. T., Dziak, J. J., Huang, L., Wagner, A. T., & Collins, L. M. Diamond, L. M. (2008). Female bisexuality from adolescence to adult- (2015). LCA Stata plugin users’ guide (version 1.2). University hood: Results from a 10-year longitudinal study. Developmental Park, PA: The Methodology Center, Penn State. Retrieved from Psychology, 44, 5–14. https​://metho​dolog​y.psu.edu/. Dworkin, S. H. (2001). Working with gay, lesbian, and bisexual clients. Laughlin, S. (2016). Gen Z goes beyond gender binaries in new innova- Journal of Clinical Psychology, 57, 671–680. tion group data. J. Walter Thompson Intelligence. Retrieved from Eisenberg, M. (2001). Diferences in sexual risk behavior between col- https​://www.jwtin​telli​gence​.com/2016/03/gen-z-goes-beyon​ lege students with same-sex and opposite-sex experience: Results d-gende​r-binar​ies-in-new-innov​ation​-group​-data/. from a national survey. Archives of Sexual Behavior, 30, 575–589. LCA Stata Plugin (Version 1.2) [Software]. (2015). University Park: England, P., Shafer, E. F., & Fogarty, A. C. K. (2008). Hooking up and The Methodology Center, Penn State. Retrieved from https://metho​ ​ forming romantic relationships on today’s college campuses. In dolog​y.psu.edu/. M. S. Kimmel & A. Aronson (Eds.), The gendered society reader Lindley, L., Nicholson, T., Kerby, M., & Lu, N. (2003). HIV/STI asso- 3 (pp. 531–547). New York: Oxford University Press. ciated risk behaviors among self-identifed lesbian, gay, bisexual, Ford, C. L., Whetten, K., Hall, S., Kaufman, J., & Thrasher, A. (2007). and transgender college students in the United States. AIDS Educa- Black sexuality, social construction, and research targeting “The tion and Prevention, 15, 413–429. down low” (“The DL”). Annals of Epidemiology, 17, 209–216. Ochs, R. (2007). What’s in a name? Why women embrace or resist bisex- Goode, E., & Haber, L. (1997). Sexual correlates of homosexual expe- ual identity. In B. A. Firestein (Ed.), Becoming visible: Counseling rience: An exploratory study of college women. Journal of Sex bisexuals across the lifespan (pp. 72–86). New York: Columbia Uni- Research, 13, 12–21. versity Press. Hagenaars, J. A., & Halman, L. C. (1989). Searching for ideal types: Ochs, R., & Deihl, M. (1992). Moving beyond binary thinking. In W. The potentialities of latent class analysis. European Sociological J. Blumenfeld (Ed.), Homophobia: How we all pay the price (pp. Review, 5, 81–96. 67–75). Boston: Beacon. Hamilton, L. (2007). Trading on heterosexuality: College women’s gen- Peplau, L. A., & Finegrhut, A. W. (2007). The close relationships of der strategies and homophobia. Gender & Society, 21, 145–172. lesbians and gay men. Annual Review of Psychology, 58, 405–424. Hightow, L. B., Leone, P., Macdonald, P., McCoy, S., Sampson, L. A., & Peterson, T. L., & Gerrity, D. A. (2006). Internalized homophobia, Kaplan, A. H. (2006). Men who have sex with men and women: A lesbian identity development, and self-esteem in undergraduate unique risk group for HIV transmission on North Carolina college women. Journal of Homosexuality, 50, 49–75. campuses. Sexually Transmitted Diseases, 33, 585–593. Phillips, L. (2005). Deconstructing “down low” discourse: The politics Hoburg, R., Konik, J., Williams, M., & Crawford, M. (2004). Bisexual- of sexuality, gender, race, AIDS, and anxiety. Journal of African ity among self-identifed heterosexual college students. Journal of American Studies, 9, 3–15. Bisexuality, 4(1–2), 25–36. Reiber, C., & Garcia, J. R. (2010). Hooking up: Gender diferences, Horowitz, J. L., & Newcomb, M. D. (2002). A multidimensional evolution, and pluralistic ignorance. Evolutionary Psychology, 8, approach to homosexual identity. Journal of Homosexuality, 42, 390–404. 1–19. Rhodes, S. D., DiClemente, R. J., Cecil, H., Hergenrather, K. C., & Yee, Hutchins, L., & Kaahumanu, L. (1991). Bi any other name: Bisexual L. J. (2002). Risk among men who have sex with men in the United people speak out. Boston: Alyson Press. States: A comparison of an interest sample and a conventional Israel, T., & Mohr, J. J. (2004). Attitudes toward bisexual women and outreach sample. AIDS Education and Prevention, 14, 141–150. men: Current research, future directions. Journal of Bisexuality, Rosario, M., Schrimshaw, E. W., Hunter, J., & Braun, L. (2006). Sexual 4(1–2), 117–134. identity development among gay, lesbians, and bisexual youths: Con- Kashubeck-West, S., & Szymanski, D. M. (2008). Risky sexual behav- sistency and change over time. Journal of Sex Research, 43, 46–58. ior in gay and bisexual men: Internalized heterosexism, sensation

1 3 Archives of Sexual Behavior

Ross, M. W., Essien, E. J., Wiliams, M. L., & Fernandez-Esquer, M. E. Szymanski, D. M., Kashubeck-West, S., & Meyer, J. (2008). Internal- (2003). Concordence between sexual behavior and sexual identity ized heterosexism: Measurement, psychological correlates, and in street outreach samples of four racial ethnic groups. Sexually research directions. The Counseling Psychologist, 36, 525–574. Transmitted Diseases, 30, 110–113. Taylor, B. (1999). “Coming out” as a life transition: Homosexual iden- Rowen, C. J., & Malcolm, J. P. (2002). Correlates of internalized homo- tity formation and its implications for health care practice. Journal phobia and homosexual identity formation in a sample of gay men. of Advanced Nursing, 30, 520–525. Journal of Homosexuality, 43, 77–92. Tein, J. Y., Coxe, S., & Cham, H. (2013). Statistical power to detect Rupp, L. J., Taylor, V., Regev-Messalem, S., Fogarty, A. C. K., & Eng- the correct number of classes in latent profle analysis. Structural land, P. (2013). Queer women in the hookup scene: Beyond the Equation Modeling: A Multidisciplinary Journal, 20, 640–657. closet? Gender & Society, 28, 212–235. Wade, L. (2017). American hookup: The new culture of sex on campus. Rust, P. C. (1992). The politics of sexual identity: Sexual attraction and New York: W.W. Norton & Co. behavior among lesbian and bisexual women. Social Problems, Walker, A. (2014a). “Our little secret”: How publicly heterosexual 39, 366–386. women make meaning from their “undercover” same-sex sexual Sanday, P. R. (2007). Fraternity gang rape: Sex, brotherhood and privi- experiences. Journal of Bisexuality, 14, 194–208. lege on campus. New York: New York University Press. Walker, A. (2014b). “I’m not a lesbian; I’m just a freak”: A pilot study Savin-Williams, R. C., & Ream, G. L. (2007). Prevalence and stability of the experiences of women in assumed-monogamous other-sex of sexual orientation components during adolescence and young unions seeking secret same-sex encounters online, their negotiation adulthood. Archives of Sexual Behavior, 36, 385–394. of sexual desire, and meaning-making of sexual identity. Sexuality Seidman, S., Meeks, C., & Traschen, F. (1999). Beyond the closet? The and Culture, 18, 911–935. changing social meaning of homosexuality in the United States. Ward, J. (2015). Not gay: Sex between straight white men. New York: Sexualities, 2, 9–34. New York University Press. Silva, T. (2017). Bud-sex: Constructing normative masculinity among Yost, M. R., & Thomas, G. D. (2012). Gender and binegativity: Men rural straight men that have sex with men. Gender & Sexuality, and women’s attitudes toward male and female bisexuals. Archives 31, 51–73. of Sexual Behavior, 41(3), 691–702. Simon, W., & Gagnon, J. H. (1986). Sexual scripts: Permanence and Young, R. M., & Meyer, I. H. (2005). The trouble with MSM and WSW: change. Archives of Sexual Behavior, 15, 97–120. Erasure of the sexual-minority person in the public health dis- Simon, W., & Gagnon, J. H. (2003). Sexual scripts: Origins, infuences, course. American Journal of Public Health, 95, 1144–1149. and changes. Qualitative Sociology, 26, 491–497. Zivony, A., & Lobel, T. (2014). The invisible stereotypes of bisexual Sophie, J. (1986). A critical examination of stage theories of lesbian men. Archives of Sexual Behavior, 43, 1165–1176. identity formation. Journal of Homosexuality, 12, 39–51.

1 3